Many electronic systems manipulate both digital and analog signals. To perform their intended function, these systems may convert analog signals into digital signals. For example, digital signal processing technology facilitates the economical and accurate transmission of either analog or digital signals to a remote receiver. In a particular application, signals in digital communications systems are transmitted as a sequence of binary pulses with the advantage that corruption of the amplitudes of these pulses by noise is, to a large extent, of no consequence. In contrast, digital video disk systems transmit and receive analog signals. In order to operate, however, these systems require circuitry to interface signals from the analog domain to signals in the digital domain so that they may perform further digital signal processing. Specifically, these systems require analog-to-digital conversion systems to interface the analog and digital domains. Advances in digital video disk systems and other related technologies indicate a need for increased sampling rates in analog-to-digital conversion systems.
Traditional analog-to-digital conversion systems use flash architectures or pipeline architectures to obtain 8-bit resolution at approximately the same conversion rate as each other. For more than 8-bit resolution, however, flash architectures are no longer feasible alternatives because they require large die areas and power dissipation. Pipeline conversion architectures attempt to reduce die areas and power requirements while increasing the sampling rate for resolutions greater than 8-bit. A March, 1992 article in the IEEE Journal of Solid-State Circuits, authored by Lewis, et al. and entitled "A 10-b 20-Msample/s Analog-to-Digital Converter," describes a particular prior art pipeline conversion system. However, these prior art systems still do not realize the optimum sampling rates attainable for a pipeline conversion system with a particular range of die areas and power dissipation.